Credit Card Default Prediction Using Machine Learning Models

Credit card default prediction is a crucial task for financial institutions. Accurately predicting the probability of default can help banks and credit card companies to better manage risk, reduce losses, and optimize their lending strategies. In this paper, a machine learning-based strategy to predict credit card default is proposed. A dataset of credit card clients from Taiwan has been utilized for the same. Various machine learning algorithms, including logistic regression, SVM, and ANN, have been used to develop predictive models. The models are evaluated under various performance metrics, including accuracy, F1 score, etc., and a set of feature scaling techniques that help to give accurate results, SVC algorithm outperforms other algorithms with an accuracy of 82.03%. Findings suggests that machine learning techniques can be used to develop accurate credit card default prediction models that can help financial institutions make informed decisions.